Application of Population-Based Linked Data to the Study of Intellectual Disability and Autism

Abstract Data linkage is the bringing together of specific datasets from different sources using demographic information on individuals within a population. In Western Australia, systematic record linkages between multiple datasets including births, deaths, and hospital morbidity have occurred since 1997. Linkage to other population datasets related to intellectual disability and autism provides an infrastructure to undertake research in this area. The purpose of this chapter is to use the Western Australian experience to describe the potential of data linkage to improve the understanding of the epidemiology of intellectual disability and autism. Information is provided on the core datasets and specific disability registers available in Western Australia. Some recently developed novel initiatives are also described. One of these involves the linkage of data from a number of other Western Australian jurisdictions such as education, child protection, and justice in addition to health. The other uses a federated approach, developed by the bioinformatics team at the Telethon Institute for Child Health Research, to enable analysis of autism risk factors from international population-based datasets without the data leaving its home base. Examples of published research examining both determinants and outcomes of intellectual disability and autism are reported as well as information on several studies, which are in the planning stages. We also include a section on some of the challenges of data management and describe some of the statistical techniques that have been used. Studies investigating sociodemographic factors, prenatal growth, alcohol exposure, and maternal physical health have identified risk factors associated with intellectual disability and autism. Analyses of outcomes for these children have shown an increased risk of hospitalizations from two to ten times that of the rest of the population. For those with Down syndrome, a quarter of their admissions occurred in the first year of life, with upper respiratory tract infections being the most common cause. Non-Aboriginal children with intellectual disability were shown to have a threefold risk of substantiated child maltreatment and Aboriginal children a twofold risk, after taking parental factors into account. Linked data studies using population data on intellectual disability and autism can help understand the complex multidimensional factors contributing to different outcomes for this particularly vulnerable and underresearched population and provide valuable information for informing policies and programs.